Diabetic retinopathy screening using single-field digital fundus photography at a district level in Costa Rica: a pilot study

Abstract

To report on our initial experience screening for diabetic retinopathy (DR) using digital fundus photography and standarized criteria. Cross-sectional prospective study of all 1327 diabetic patients referred to the Ophthalmology Department of the Clorito Picado Clinic in San José, Costa Rica between April 2008 and March 2009. Patients were screened with a single-field 45-degree digital color retinal image centered on the fovea. The main outcome measures were prevalence of DR, waiting time for a fundus photograph appointment and the number of patients that needed to be evaluated by an ophthalmologist. In this population of 1327 patients, 85% did not have any retinopathy. The prevalence of DR was 15%. Mild non-proliferative DR (NPDR) was seen in 5.8% (70/1215), moderate NPDR in 3.9% (47/1215), severe NPDR in 4.7% (57/1215) and proliferative DR in 0.6% (8/1215). Macular edema was present in 2.6% (32) patients. 13.8% of patients required pupillary dilation for an adequate image to be taken and 8.8% of the images were unreadable. The average waiting time for a fundus photograph appointment was 49.1 days. A total of 13.8% (183/1327) patients were referred to the ophthalmologist. Of these referrals, 61.2% (112/183) were because of unreadable images and 38.8% (71/183) were referred for further ophthalmic treatment. Digital fundus photographic screening for DR appears to be a viable option for the improvement of screening for DR in Costa Rica.

Addressing risk factors, screening, and preventative treatment for diabetic retinopathy in developing countries: a review

“…The majority of the Asian studies were from India ( n = 9) and China ( n = 6), the two countries with the largest diabetes populations worldwide. When diabetes type was known and recorded, the vast majority of patients had type 2 DM (T2DM) (81.6%–99.4%).…”

“…Referral rates from the screening programs ranged from 5.9% to 29%, depending on both the prevalence of STDR (5%–27.6%), as well as other criteria for referral. In particular, for two programs using fundus photograph screening, ungradable images made up 61.2% and 65% of referrals, respectively. Besides ungradable images, PDR, STDR and DME were the most common reasons for referral.…”

“…Common reasons for ungradable images included severe cataract, as well as corneal or vitreous opacities . Programs with lower rates of ungradable images had quality control mechanisms embedded in the image‐taking process, either automated or based on expertise of the photographer.…”

“…One mobile van documented in Tamil Nadu, India was able to screen at least 150 patients per day, with one ophthalmologist making final diagnoses from the base hospital . One study in Costa Rica estimated that a photographer could photograph 6 patients per hour, and an ophthalmologist could read photos for 20 patients an hour . This approach is seen as especially important in developing countries, where numbers of ophthalmologists average 29 per 1 million population .…”

“…Common reasons for ungradable images included severe cataract, as well as corneal or vitreous opacities. 39,44 Programs with lower rates of ungradable images had quality control mechanisms embedded in the image-taking process, either automated or based on expertise of the photographer. For example, multiple studies suggested that NMC may need to be improved for non-Caucasian ethnic groups, perhaps because machines have not yet been trained on dark irises.…”